Beruflich Dokumente
Kultur Dokumente
2018
1
Department of Mechanical Engineering,
Covenant University, Ota, Ogun State, Nigeria
2
Department of Mechanical Engineering,
University of Lagos, Akoka, Lagos, Nigeria.
Abstract─ The fouling of heat exchangers in the oil and gas industry is not
only a recurring challenge in refineries but it is also becoming a challenge in
crude oil production and treatment facilities where heat exchangers are
deployed to cool the crude oil temperature downstream of stabilization
system prior to storage. A predictive mitigation approach to addressing
fouling in heat exchangers remains the most viable option to avoid
production train capacity limitations, unplanned shutdown and attendant loss
of revenue. Considerable researches have been carried out which lead to the
development of models used for predicting fouling resistances in shell and
tube heat exchangers. However, this study focused on evaluation of the
performance of a plate and frame heat exchanger utilized in cooling of crude
oil prior to storage in a cargo tank for export. It also developed a fouling
resistance suitable for forecasting the exchanger performance and predicting
the maintenance management scheme. The data employed were continuously
measured for three years and employed for the analysis. It found that the
lower the hot stream approach temperature, the more the fouling resistance.
In addition, the work validates that as the fouling resistance increases with
time, the efficiency of the plate and frame heat exchanger diminishes.
Keywords: Fouling Models, Fouling Resistance, Energy System, Heat
Exchanger, Scaling
72
Stephen K. Ogbonnaya, et al CJET (2018) 2(2) 72-90
73
Stephen K. Ogbonnaya, et al CJET (2018) 2(2) 72-90
74
Stephen K. Ogbonnaya, et al CJET (2018) 2(2) 72-90
75
Stephen K. Ogbonnaya, et al CJET (2018) 2(2) 72-90
Duty 14.75MW
Heat Transfer Rate 2325W/m2-oC
Mean Temperature Difference 26.28oC
The analysis of a heat exchanger can temperature of the cooling water and
be performed by either the Number of pressure drop across the heat
Transfer Units (NTU) method or the exchanger. These operating
log mean temperature difference parameters were utilized with the
(LMTD) method [17]. The LMTD HTRI Xchanger Suite to generate the
method is applied to problems for cooling water outlet temperature,
which the fluid flow rates and inlet duty of the plate and frame, overall
temperatures, as well as a desired heat transfer coefficient and fouling
outlet temperature, are prescribed. resistance at various operating points.
Therefore, a heat exchanger modeling The values of the various operating
application that utilizes the LMTD data were detected by various
approach was used. instrumentation devices like the
Data Gathering turbine flow meters, temperature
The multi-year operation data of the transmitters and pressure gauges. The
plate and frame heat exchanger values of some parameters like the
employed for the study were sourced flow rates and temperatures are
from an oil and gas firm in Nigeria. automatically transmitted to a system
They include the mass flow rate of of record (SoR) called the
the hot crude oil, mass flow rate of Exaquantum which both stores the
the cooling water, inlet temperature data for several years as well as
of the hot crude oil, outlet provides real time values of the
temperature of the hot crude oil, inlet operating parameters at any point in
time. Some other operating data like
76
Stephen K. Ogbonnaya, et al CJET (2018) 2(2) 72-90
the pressure drop across the heat coefficient, Us, to clean overall heat
exchanger was manually recorded by transfer coefficient, Ucl, (Us/Ucl) and
operating personnel by reading the the Fouling Resistance (Rf) were
pressure gauges on the inlet and exit monitored. Variations in Us/Ucl
lines of the unit. defines the change in heat transfer
A total of over seventeen thousand capacity with time and hence the heat
data points covering the periods transfer efficiency. It was noted that
between 2013 and 2015 were collated tracking Us/Ucl alone can be
from the datasheet for each of the misleading, because it is a hyperbolic
operating parameters making it a total function and hence was more reliable
of 68000 data points. In addition, to monitor the fouling resistance, Rf,
1200 manually recorded data points directly using the equation:
were also used. Upon reviewing the (m2 °C/W) (3)
data, some data were found to be
incomplete for various reasons which Heat Transfer Efficiency
included instrumentation (device) The heat exchanger performance is
malfunction, human errors, typically measured in terms of heat
measurement error and emergency transfer efficiency and hydraulic
shutdown of the facility. Based on capacity [18]. In this work, the heat
this, the data were treated in two transfer efficiencies of the heat
ways namely: exchanger over a period was
i. Incomplete data: This results when determined, according to [18], as the
the system did not capture the ratio of service overall heat transfer
values of all the key operating coefficient, Us, to clean overall heat
parameters at certain days. Due to transfer coefficient, Ucl, both of
the criticality of data which are defined in equations 4 and
completeness, days with 5
incomplete data were removed. (4)
ii. Outliers: Some data were provided
(5)
but were found to be significantly
distant from other measurements
recorded and so were excluded where: Us = Service overall heat
from data set. The outliers were transfer coefficient, (W/m2 °C), Q =
evaluated statistically using the Rate of heat transferred, (W), A =
Three Sigma Method denoted by Outside area of tubes, (m2), EMTD =
equation (2) and this represented Effective mean temperature
0.07% of the overall data. difference, (°C), Ucl = Clean overall
heat transfer coefficient, (W/m2 °C),
3 (2)
hi = Inside film heat transfer
Development of the Heat coefficient based on inside area,
Exchanger Model (W/m2 °C), Do = Outside diameter of
The model development was carried tubes,( m2), Di = Inside diameter of
out using the Heat Transfer Research tubes, (m2), ho = Outside film heat
Inc (HTRI) Xchanger Suite and transfer coefficient based on outside
Parametric Study Spreadsheet. The area, (W/m2 °C), Rw = Wall
performance parameters such as the resistance, (m2 °C/W)
ratio of service overall heat transfer
77
Stephen K. Ogbonnaya, et al CJET (2018) 2(2) 72-90
n 2
RMSE ( H m H cal ) / k
(6)
i 1
n
MBE ( H m H cal ) / k (7)
i 1
k H m H cal
100
MPE i 1 Hm (8)
k
k
H Hm
2
cal
(9)
COE 1 i 1
k
H Hm
2
cal
i 1
k H -H
i 1
m
Hm
cal
100
MAPE
k (10)
n
MABE ( H m H cal ) / k
i 1 (11)
Where: k = number of data points, H m = mean of all the fouling resistance
obtained from experiment.
III. Results and Analysis design parameters are displayed in
Performance of the Heat Exchanger Table 2. The Table shows that the
The accuracy of the modeling and results from HTRI are very close to
simulation carried out using the HTRI those of manufacturer’s datasheet of
was first determined by checking the the in-service plate and frame heat
convergence of the results to the exchanger employed for the study.
original design parameters for the The degree of accuracy of the results
plate and frame heat exchanger. The therefore provides the basis for the
comparison of the results from the subsequently performance study
HTRI to that of the manufacturer’s using the HTRI.
78
Stephen K. Ogbonnaya, et al CJET (2018) 2(2) 72-90
Table 2: Comparison of results from simulation with that from manufacturer’s datasheet
S/No. Parameter Data Sheet Value Modelling results
1 Actual Heat Transfer 2325W/m2-k 2204.8W/m2-k
Coefficient
2 Required Heat Transfer 2627W/m2-K 2282.4W/m2-k
Coefficient
3 Duty 14.7MW 14.53MW
4 Area 241.14m2 241.27m2
5 Effective Mean Transfer 26.28oC 26.4oC
Difference
6 Quantity of Plates 267 267
The heat exchanger and the Frame in series representation are shown in Figures 1 and 2.
81.1 C 43.0 C
0.00000 0.00000
22.2 C 48.3 C
0.00000 0.00000
ID 100
14.5258 MegaWatts
%OD -3.4
Figure 2: Frame in Series Representation of Heat Exchanger
79
Stephen K. Ogbonnaya, et al CJET (2018) 2(2) 72-90
The variations of the heat transfer efficiency over time are shown in figure 3 (a-c)
Figure 3a: Heat Transfer Efficiency of the heat exchanger (January - December 2013)
Figure 3b: Heat Transfer Efficiency of the heat exchanger (January - December 2014)
Figure 3c: Heat Transfer Efficiency of the heat exchanger (January - December 2015)
80
From Figure 3 (a – c), the heat end of the year. The dynamic nature
transfer efficiency for each year of the fluid properties and the process
studied tends to reduce over time. is essentially responsible for the
The downward trend of the heat behaviour of the efficiencies from
transfer efficiency can be year to year. The temperature
extrapolated to estimate when the difference was noted as responsible
efficiency drop would become a for the disparity in heat transfer
concern, hence facilitating early efficiency behaviour over the years of
planning for possible maintenance study. The approach temperature
action. The decline in heat transfer (~average of 55oC) of the crude was
efficiency was more pronouced in lowest in 2013 due to the then
2013 than in 2014 and 2015. Based reservoir conditions unlike those of
on the field data,the heat transfer 2014 (~average of 58oC) and 2015
efficiency was 85% as at the (~average of 62oC). Lower crude
beginning of 2013; but it dropped to approach temperature imply lower
72% by the end of the year. In 2014, temperature difference between the
the heat transfer efficiency dropped crude and cooling water and vice
from 92% at the beginning of the year versa.
to 84% at the end of the year while in Moreoever, the changes in the fouling
2015 it was 92% at the beginning of resistance Rf over time are presented
the year and declined to 83% at the in figure 4 (a-c):
Figure 4a: Progression Fouling Resistance of the heat exchanger for the period of January
to December 2013
Figure 4b: Progression Fouling Resistance of the heat exchanger for the period of January
to December 2014
81
Stephen K. Ogbonnaya, et al CJET (2018) 2(2) 72-90
Figure 4c: Progression Fouling Resistance of the heat exchanger for the period of January
to December 2015
54oC. In 2014, the fouling resistance
The figure 4 (a – c) shows the fouling ranged from 0.000005 m2-K/W at
resistances were seen to increase over 63oC to 0.00020 m2-K/W at 59oC. In
time for each year. The dynamics of 2015, it ranged from 0.000005m2-
the operating conditions such as K/W at 62oC to 0.000165 m2-K/W at
changes in flow rate and variations in 56oC. Hence, it was observed that
feed temperature were responsible for fouling resistance tends to increase
the fluctuations in data points. with lower feed temperature in a plate
According to [19], fouling of heat and frame heat exchanger used for
transfer surfaces is a complex process crude oil services. However, an
which involves many parameters with understanding of the performance of
poorly understood interactions. the heat exchanger in relation to its
Hence, it could result in highly fouling resistance is essential in order
unstable processes with frequent to forecast its impact on the
significant fluctuation, if the variation performance and ascertain the best
of the heat transfer coefficient or maintenance on the unit. This will
fouling resistance was observed with require the development of a
time. The increase in fouling predictive model that explains the
resistance over time was more in phenomenon of fouling resistance
2013 than in 2014 and 2015 with input parameters. Figures 5a and
respectively. In 2013, the fouling 5b show the trends of the combined
resistance ranged from 0.000035 m2- heat transfer efficiency and fouling
K/W at 61oC to 0.000320 m2-K/W at resistance with time respectively.
82
Figure 5a: Progression of Combined (Multi-Year) Heat Transfer Efficiency
83
Stephen K. Ogbonnaya, et al CJET (2018) 2(2) 72-90
(15)
(16)
where: = Cooling Water Inlet e, f, g are correlation coefficients.
Temperature (0C), = Crude Oil The values of the correlation
Inlet Temperature (0C), = Crude coefficients as determined from Eqs.
0 12 to 16 are shown in Table 3, while
Oil Outlet Temperature ( C), =
those of the error and performance
Crude Oil Mass Flow rate (kg/s), t =
analysis for the different models are
Precise Day of Operation within a
shown in Table 4.
specified period (day), = Fouling
Resistance (m2-K/oC), and a, b, c, d,
Table 3: Regression Model Constants
Coefficients MODEL 1 MODEL 2 MODEL 3 MODEL 4 MODEL 5
a -2.9563E-05 -1.49156E-05 -2.7E-06 -2.955E-05 -3E-05
b 4.83066E-05 -4.42831E-06 -9.1E-08 4.82E-05 4.8E-05
c -1.96825E-05 -2.57434E-06 0.00058 -1.965E-05 -2E-05
d -3.26233E-06 -5.62769E-08 0 -3.243E-06 -9.8E-09
e -2.00747E-09 0.001100191 0 8.186E-12 -1E-09
f 0.000682127 0 0 -8.819E-09 0.000424
g 0 0 0 0.0006811 0
84
Stephen K. Ogbonnaya, et al CJET (2018) 2(2) 72-90
Figure 6a: Comparison between measured Fouling Resistance with calculated Fouling
Resistance using Equation (12)
Figure 6b: Comparison between measured Fouling Resistance with calculated Fouling
Resistance using Equation (13)
85
Figure 6c: Comparison between measured Fouling Resistance with calculated Fouling
Resistance using Equation (14)
Figure 6d: Comparison between measured Fouling Resistance with calculated Fouling
Resistance using Equation (15)
86
Figure 6e: Comparison between measured Fouling Resistance with calculated Fouling
Resistance using Equation (16)
Model Evaluation
Based on the model analysis, accurate model while equation (18)
equation (17) presents the most closely followed
R f 4.82T 2.96Twi 1.97Toi 0.324 x 101 M o 8.19 x 107 t 2 8.82 X 104 t 0.0006811 105
(17)
R f 4.83T – 2.96Twi 1.97Toi 0.326 x 101 M o – 2.01 x 104 t 68.21 105
(18)
Based on the aforementioned, type of intervention will be required.
equations (17) and (18) predict the Table 5 therefore, provides the
fouling resistance of a plate and predicted fouling resistances of a
frame heat exchanger, enabling the plate and frame heat exchanger for a
forecast of a time when a period (days) using the best model
maintenance program or any other above at operating conditions.
For instance, if a plate and frame stream temperature, the lower the hot
heat exchanger is installed in January stream approach temperature, the
2016, the fouling resistance would more the fouling resistance. Hence
have increased to 6.43 x 10-5 m2 there is a relationship between the
°C/W by January 2021. Having this inlet temperature of the hot process
knowledge will enable the operator stream and the fouling resistance.
know that heat transfer efficiency Moreover, due to the complexity of
will be impacted in five years’ time fouling phenomenon, further work
and then adequate plan can be put in may be required to develop the
place to clean the heat exchanger or fouling rate models of plate and
replace with a spare unit. frame heat exchangers for various
IV Conclusion crudes and crude blends with
The study focused on investigation different thermophysical properties
of the performance of a plate and and chemical compositions.
frame heat exchanger employed in an Furthermore, computational fluid
oil and gas production/treatment dynamics can be applied to the study
facility of upstream sector with to be able to incorporate other fluid
operational data. It developed properties and also establish the
multivariate regression models that pathway to easy fouling prediction
can be used to predict required and maintenance set-up programme.
maintenance plan and forecast the Acknowledgement
economic implication of fouling on The authors thank the management
the heat exchanger. Apart from the of Covenant University for the
results already highlighted, the study support received towards the
also found that at a constant cold complete execution of this project.
88
Covenant Journal of Engineering Technology (CJET). Vol.1 No.1, Feb. 2018 (Special Edition)
[5] Morimoto N., Nakamura M., [10] Ho, T. C. (2016). A study of
Yamane Y. and Inokuchi H. crude oil fouling propensity.
(2011). Development of a Hydrocarbon Conversion
Conceptual Model of Technologies, Bridgewater, NJ
Crystallization in a Heat 08807, United States.
Exchanger. Proc. of International Journal of Heat
International Conference on and Mass Transfer 95 (2016)
Heat Exchanger Fouling and 62–68
Cleaning - June 05 - 10, 2011, [11] Lane, M. R. and Harris J. S.
Crete Island, Greece (2015). Impact of Mixing on
[6] Polley G.T., Morales A. F. and Crude Oil Fouling Tests.
Pugh S.J. (2009). Design of Proceedings of International
Shell-And-Tube Heat Conference on Heat Exchanger
Exchangers to Achieve a Fouling and Cleaning - 2015
Specified Operating Period in (Peer-reviewed) June 07 - 12,
Refinery Pre-Heat Trains. Proc. 2015, Enfield (Dublin), Ireland
of International Conference on 44 [12] Bennett, C. A. (2016). A
Heat Exchanger Fouling and Theory Describing
Cleaning VIII - 2009 June 14- Sedimentation Particulate
19, Schladming, Austria Fouling Thresholds Inside Heat
[7] Costa, A.L.H., Tavares, V.B.G., Exchanger Tubes. Heat Transfer
Borges J.L., Queiroz, E.M., Engineering Volume 37, 2016 -
Pessoa, F.L.P., Liporace, F.S. Issue 5
and Oliveira, S.G. (2011). [13] Emani, S., Ramasamy, M.,
Parameter Estimation of Zilati, K., Shaari, B. K. (2016).
Fouling Models in Crude Effect of Shear Stress on Crude
Preheat Trains. Proceedings of Oil Fouling in a Heat
International Conference on Exchanger Tube Through CFD
Heat Exchanger Fouling and Simulations. 4th International
Cleaning - 2011 (Peer- Conference on Process
reviewed) June 05 - 10, 2011, Engineering and Advanced
Crete Island, Greece Materials Procedia Engineering
[8] Saleh, Z., Sheikholeslami, R. and 148 (2016) 1058 – 1065Nelson,
Watkinson, A.P., (2003). G. L. (1995). Adhesion, in paint
Fouling characteristics of a light and coating testing manual,
Australian crude oil, Proc. Heat (14thed), In J. V. Koleske,
Exchanger Fouling and Philadelphia P.A: ASTM, 44.
Cleaning: Fundamentals and [14] Wilson, D.I., Ishiyama, E.M.,
Applications, 226 – 233, Santa Polley, G.T, (2015). Twenty
Fe, NM. Years of Ebert And Panchal –
[9] Shetty, N., Deshannavar, U. B., What Next? Proceedings of
Marappagounder, R., Pendyala, International Conference on
R., (2016). Improved threshold Heat Exchanger Fouling and
fouling models for crude oils. Cleaning - 2015 (Peer-
Energy Volume 111, 15 reviewed) June 07 - 12, 2015,
September 2016, Pages 453– Enfield (Dublin), Ireland
467
89
Covenant Journal of Engineering Technology (CJET). Vol.1 No.1, Feb. 2018 (Special Edition)
[15] Yang, J., Serratos, M. G. J., [17] Çengel Y.A. and Ghajar A. J.
Fari-Arole, D. S., Müller, E. A., (2015). Heat and Mass
Matar, O. K. (2015). Crude Oil Transfer: Fundamentals &
Fouling: Fluid Dynamics, Applications, Fifth Edition
Reactions and Phase Change. Published by McGraw-Hill
Procedia IUTAM 15 (2015) 186 Education.
– 193. [18] Al-Janabi A., Esawy M.,
[16] Joshi, H. M., Shilpi, N. B., and Malayeri M.R. and Müller-
Agarwal, A.(2009). Relate Steinhagen H. (2009).
Crude Oil Fouling Research to Consideration of Dynamic
Field Fouling Observations. Uncertainty in Fouling
Proceedings of International Experimentation. Proc. of
Conference on Heat Exchanger International Conference on
Fouling and Cleaning VIII - Heat Exchanger Fouling and
2009 (Peer-reviewed) June 14- Cleaning VIII - June 14-19,
19, 2009, Schladming, Austria 2009, Schladming, Austria
90